A Variable Strength Interaction Test Suites Generation Strategy using Sewing Training Optimization

Abstract

This paper reports on a study in which a new model for the development of Variable Strength (VS) interaction test suits, identified as Sewing Training Based Optimization (STBO).  This approach is based on the principle of how tailor apprentices learn through a stage of imitation, a stage of practice, and a stage of evaluation, which together bring about a balance between wide exploration and focused search. To evaluate its performance in terms of test suite size reduction and scalability, STBO was applied to commonly used base models. The results indicate that STBO produced 26 optimal solutions, outperforming nearly 90% of existing methods. Consistent superiority over traditional algorithms was observed, while performance remained comparable to that of more complex metaheuristic approaches. In addition, a real-time software system case study was conducted, in which STBO demonstrated effective fault detection capability, thereby indicating that it is a reliable and adaptive strategy for software testing.